There is considerable interest in systematically analyzing the human plasma proteome to identify novel biomarkers that can be used for improved early diag
* Originally published in Proteomics 2005, 13, 3329-3342
nosis of a wide range of diseases. Plasma or serum is easily and widely collected and its proteome contains thousands of proteins including proteins secreted or shed by most cells and tissues as well as proteins that leak into the blood from damaged tissue . The presence or change in concentration of blood proteins is likely to reflect the state of health of an individual. A number of proteins discovered through targeted studies are currently being used as diagnostic markers for diseases such as acute myocardial infarction (creatine kinase MB, myoglobin, and troponin T ), prostate cancer (prostate-specific antigen ), and ovarian cancer (CA125 ). However, it is likely that the blood contains many additional disease biomarkers that will have greater diagnostic value than the handful of biomarkers discovered.
While the human plasma proteome potentially contains many different important biomarkers for most human diseases, several factors make it difficult to characterize. Plasma proteins are present in a very wide dynamic range, varying by a factor of at least 1010 in abundance, and many of these proteins have a high degree of heterogeneous PTMs . The ability to identify low-abundance plasma proteins is particularly severely limited by several major proteins that are present at >1 mg/mL. For example, albumin together with immunoglobulins contributes to more than 80% of the total plasma proteins at about 40 and 12 mg/mL, respectively [1, 5]. In contrast, many bioactive proteins and potential biomarkers of disease are low-abundance proteins that are typically found at ng/mL - pg/mL levels or less.
The strategies that have been most frequently used to overcome the dynamic range problem of plasma proteins are to fractionate the plasma proteome into smaller subsets, and/or to deplete one or more of the major proteins, particularly albumin and immunoglobulins [5-12]. Numerous dye-based and immu-noaffinity methods for major protein depletion have been described and are available commercially. Immunoaffinity methods are preferred, as they provide the most efficient depletion of targeted major proteins with reduced nonspecific binding of other proteins [7-9]. Alternatively, albumin can be efficiently separated based on its pi by microscale solution IEF (MicroSol-IEF) into a single fraction [10, 11]. Both major protein depletion and MicroSol-IEF methods have resulted in increased detection of lower abundance proteins when analyzed by 2-DE [8, 9, 11]. While removal of major proteins is beneficial, multiple orthogonal fractionation steps have been used to further facilitate detection of low-abundance proteins [8, 12].
A popular alternative to 2-DE is the shotgun or multidimension protein identification technology (MudPIT) approach which involves proteolytically digesting complex protein mixtures into peptides that are further subjected to multidimensional separations prior to analysis by ESI-MS/MS . The most common form of multidimensional separations involves strong cation exchange (SCX) chromatography followed by RP-LC [12, 14, 15]. Alternate peptide separation strategies, such as ampholyte-free liquid-phase IEF  and CZE , have also been used in the analysis of human serum proteome.
Compared to 2-DE, the MudPIT approach has the potential of higher throughput and is capable of identifying more proteins from the plasma proteome. In a 2-DE study, 325 proteins were identified from human serum after 3-D fractionation using immunodepletion of nine abundant proteins, anion-exchange, and SEC . In comparison, 490 proteins were identified with the MudPIT technique using immunoglobulin depletion and 2-D peptide separations by SCX and RP-LC . While the 2-DE technology is relatively mature, the MudPIT method is constantly improving due to technological advances mainly to the LC and MS components of the system. In a recent study using ultra-high-performance SCX/RP-LC coupled to MS/MS, at least 800 proteins (depending on the criteria used) were identified from human plasma proteome . These proteins were identified without prior depletion of major proteins, indicating that the improvement to the LC system and the longer gradient used were capable of overcoming the dynamic range problem of the plasma proteome to a certain degree. However, since immunoglobulins, which contain highly variable regions, were not depleted in the study, many of the proteins identified (up to 38%) belong to the immunoglobulin group .
The total number of proteins in the human plasma proteome is unknown but has been estimated to contain up to 10 000 proteins . A recent analysis of the human plasma proteome by combining four separate sources of protein identification, including a 2-DE and two separate MudPIT experiments, has resulted in a conservative nonredundant list of 1175 proteins . Interestingly, only 46 proteins are common to all four sources. This indicates that current methodologies cannot consistently provide comprehensive coverage of the human plasma proteome. Clearly, further reduction in the complexity of the human plasma proteome by additional more efficient fractionation steps is required to effectively mine the lower abundance proteins that have potential to be the next generation of disease biomarkers. Realizing the need for better methodology to analyze the human plasma proteome, HUPO has established the Plasma Proteome Project (PPP), and one of its aims is to determine the best technology platform for comprehensive profiling ofthe human plasma and serum proteomes .
As a participant of the HUPO PPP, in this report we describe a novel 4-D separation strategy to analyze the human plasma and serum proteomes that combines many of the benefits of 2-DE and MudPIT approaches. This strategy, termed protein array pixelation, consists of three sequential protein fractionation methods (major protein depletion, MicroSol-IEF fractionation, and SDS-PAGE). The result is a 2-D array of pixels or gel slices that is conceptually equivalent to a low-resolution 2-D gel. That is, each pixel in the array contains a group of proteins in a gel slice with a known pi and molecular weight (MW) range. Each pixel is then digested with trypsin followed by RP-LC peptide separation prior to ESI-MS/MS analysis. Using HUPO plasma and serum samples, we demonstrate that the protein array pixelation strategy is a highly sensitive method capable of detecting proteins that differ in abundance up to nine orders ofmagnitude.
138 6 A novel four-dimensional strategy combining protein and peptide separation methods 6.2
Was this article helpful?